CodeIsAbstract commited on
Commit
36984eb
·
verified ·
1 Parent(s): e0bbc82

Training in progress, step 500, checkpoint

Browse files
last-checkpoint/config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "HybridTimeScaleLM"
4
+ ],
5
+ "bos_token_id": 1,
6
+ "chunk_size": 128,
7
+ "dropout": 0.05,
8
+ "dtype": "bfloat16",
9
+ "eos_token_id": 2,
10
+ "latent_dim": 768,
11
+ "layer_types": [
12
+ "linear",
13
+ "linear",
14
+ "linear",
15
+ "softmax",
16
+ "linear",
17
+ "linear",
18
+ "linear",
19
+ "softmax",
20
+ "linear",
21
+ "linear",
22
+ "linear",
23
+ "softmax"
24
+ ],
25
+ "model_type": "hybrid_fourier_lm",
26
+ "num_layers": 12,
27
+ "num_modes": 64,
28
+ "pad_token_id": 2,
29
+ "tie_word_embeddings": true,
30
+ "time_scale": 128.0,
31
+ "transformers_version": "5.14.1",
32
+ "use_cache": false,
33
+ "vocab_size": 32768
34
+ }
last-checkpoint/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "output_attentions": false,
6
+ "output_hidden_states": false,
7
+ "pad_token_id": 2,
8
+ "transformers_version": "5.14.1"
9
+ }
last-checkpoint/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cad8695f9611fd72a1cb7da7c2dc7722ffd94491b413b08931297f19a10ee048
3
+ size 234681136
last-checkpoint/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f17b41e804bbacf24053241af7cfe65159104d0213019d3237630b978e42eda
3
+ size 469516363
last-checkpoint/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76cd281a3ce995fc434ac757f1094066d955dd6b7188cf148f0dd1523cb9cd80
3
+ size 14645
last-checkpoint/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e63258c783b6bbb756d9067a003d6c6c729901571b495c03dd6472af6b5a7f7b
3
+ size 1465
last-checkpoint/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
last-checkpoint/tokenizer_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<s>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "</s>",
7
+ "is_local": false,
8
+ "legacy": false,
9
+ "local_files_only": false,
10
+ "model_max_length": 1000000000,
11
+ "pad_token": "</s>",
12
+ "sp_model_kwargs": {},
13
+ "spaces_between_special_tokens": false,
14
+ "tokenizer_class": "TokenizersBackend",
15
+ "unk_token": "<unk>",
16
+ "use_default_system_prompt": false
17
+ }
last-checkpoint/trainer_state.json ADDED
@@ -0,0 +1,424 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.07142857142857142,
6
+ "eval_steps": 100,
7
+ "global_step": 500,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0014285714285714286,
14
+ "grad_norm": 0.060791015625,
15
+ "learning_rate": 8.571428571428572e-07,
16
+ "loss": 2.637472915649414,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.002857142857142857,
21
+ "grad_norm": 0.053955078125,
22
+ "learning_rate": 1.8095238095238097e-06,
23
+ "loss": 2.6134210586547852,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.004285714285714286,
28
+ "grad_norm": 0.05126953125,
29
+ "learning_rate": 2.7619047619047625e-06,
30
+ "loss": 2.617936134338379,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.005714285714285714,
35
+ "grad_norm": 0.052978515625,
36
+ "learning_rate": 3.7142857142857146e-06,
37
+ "loss": 2.6247318267822264,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.007142857142857143,
42
+ "grad_norm": 0.049560546875,
43
+ "learning_rate": 4.666666666666667e-06,
44
+ "loss": 2.6199132919311525,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.008571428571428572,
49
+ "grad_norm": 0.049560546875,
50
+ "learning_rate": 5.619047619047619e-06,
51
+ "loss": 2.6097543716430662,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.01,
56
+ "grad_norm": 0.10986328125,
57
+ "learning_rate": 6.571428571428572e-06,
58
+ "loss": 2.6390161514282227,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.011428571428571429,
63
+ "grad_norm": 0.3671875,
64
+ "learning_rate": 7.523809523809524e-06,
65
+ "loss": 2.6347326278686523,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.012857142857142857,
70
+ "grad_norm": 0.796875,
71
+ "learning_rate": 8.476190476190477e-06,
72
+ "loss": 2.6170269012451173,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.014285714285714285,
77
+ "grad_norm": 5.34375,
78
+ "learning_rate": 9.42857142857143e-06,
79
+ "loss": 2.62125186920166,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.014285714285714285,
84
+ "eval_loss": 3.02799129486084,
85
+ "eval_runtime": 4.4776,
86
+ "eval_samples_per_second": 66.554,
87
+ "eval_steps_per_second": 3.797,
88
+ "step": 100
89
+ },
90
+ {
91
+ "epoch": 0.015714285714285715,
92
+ "grad_norm": 0.4140625,
93
+ "learning_rate": 1.0380952380952383e-05,
94
+ "loss": 2.6266862869262697,
95
+ "step": 110
96
+ },
97
+ {
98
+ "epoch": 0.017142857142857144,
99
+ "grad_norm": 0.0634765625,
100
+ "learning_rate": 1.1333333333333334e-05,
101
+ "loss": 2.646818161010742,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 0.018571428571428572,
106
+ "grad_norm": 0.3671875,
107
+ "learning_rate": 1.2285714285714288e-05,
108
+ "loss": 2.616552543640137,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 0.02,
113
+ "grad_norm": 0.04638671875,
114
+ "learning_rate": 1.3238095238095238e-05,
115
+ "loss": 2.619216728210449,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 0.02142857142857143,
120
+ "grad_norm": 0.0498046875,
121
+ "learning_rate": 1.4190476190476192e-05,
122
+ "loss": 2.652239990234375,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 0.022857142857142857,
127
+ "grad_norm": 0.057373046875,
128
+ "learning_rate": 1.5142857142857144e-05,
129
+ "loss": 2.5990116119384767,
130
+ "step": 160
131
+ },
132
+ {
133
+ "epoch": 0.024285714285714285,
134
+ "grad_norm": 0.4140625,
135
+ "learning_rate": 1.6095238095238096e-05,
136
+ "loss": 2.635771369934082,
137
+ "step": 170
138
+ },
139
+ {
140
+ "epoch": 0.025714285714285714,
141
+ "grad_norm": 0.0517578125,
142
+ "learning_rate": 1.704761904761905e-05,
143
+ "loss": 2.617422103881836,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 0.027142857142857142,
148
+ "grad_norm": 0.064453125,
149
+ "learning_rate": 1.8e-05,
150
+ "loss": 2.64025764465332,
151
+ "step": 190
152
+ },
153
+ {
154
+ "epoch": 0.02857142857142857,
155
+ "grad_norm": 0.046875,
156
+ "learning_rate": 1.8952380952380953e-05,
157
+ "loss": 2.6089372634887695,
158
+ "step": 200
159
+ },
160
+ {
161
+ "epoch": 0.02857142857142857,
162
+ "eval_loss": 3.0282089710235596,
163
+ "eval_runtime": 4.3641,
164
+ "eval_samples_per_second": 68.284,
165
+ "eval_steps_per_second": 3.895,
166
+ "step": 200
167
+ },
168
+ {
169
+ "epoch": 0.03,
170
+ "grad_norm": 0.0537109375,
171
+ "learning_rate": 1.9904761904761908e-05,
172
+ "loss": 2.6311504364013674,
173
+ "step": 210
174
+ },
175
+ {
176
+ "epoch": 0.03142857142857143,
177
+ "grad_norm": 0.05078125,
178
+ "learning_rate": 2e-05,
179
+ "loss": 2.6472368240356445,
180
+ "step": 220
181
+ },
182
+ {
183
+ "epoch": 0.032857142857142856,
184
+ "grad_norm": 0.05322265625,
185
+ "learning_rate": 2e-05,
186
+ "loss": 2.6085578918457033,
187
+ "step": 230
188
+ },
189
+ {
190
+ "epoch": 0.03428571428571429,
191
+ "grad_norm": 0.046630859375,
192
+ "learning_rate": 2e-05,
193
+ "loss": 2.6177978515625,
194
+ "step": 240
195
+ },
196
+ {
197
+ "epoch": 0.03571428571428571,
198
+ "grad_norm": 0.06298828125,
199
+ "learning_rate": 2e-05,
200
+ "loss": 2.6038957595825196,
201
+ "step": 250
202
+ },
203
+ {
204
+ "epoch": 0.037142857142857144,
205
+ "grad_norm": 0.048095703125,
206
+ "learning_rate": 2e-05,
207
+ "loss": 2.6111982345581053,
208
+ "step": 260
209
+ },
210
+ {
211
+ "epoch": 0.03857142857142857,
212
+ "grad_norm": 0.04833984375,
213
+ "learning_rate": 2e-05,
214
+ "loss": 2.600232696533203,
215
+ "step": 270
216
+ },
217
+ {
218
+ "epoch": 0.04,
219
+ "grad_norm": 0.046630859375,
220
+ "learning_rate": 2e-05,
221
+ "loss": 2.591439056396484,
222
+ "step": 280
223
+ },
224
+ {
225
+ "epoch": 0.041428571428571426,
226
+ "grad_norm": 0.04833984375,
227
+ "learning_rate": 2e-05,
228
+ "loss": 2.6285400390625,
229
+ "step": 290
230
+ },
231
+ {
232
+ "epoch": 0.04285714285714286,
233
+ "grad_norm": 40.0,
234
+ "learning_rate": 2e-05,
235
+ "loss": 2.6076997756958007,
236
+ "step": 300
237
+ },
238
+ {
239
+ "epoch": 0.04285714285714286,
240
+ "eval_loss": 3.02835750579834,
241
+ "eval_runtime": 4.3773,
242
+ "eval_samples_per_second": 68.079,
243
+ "eval_steps_per_second": 3.884,
244
+ "step": 300
245
+ },
246
+ {
247
+ "epoch": 0.04428571428571428,
248
+ "grad_norm": 0.049072265625,
249
+ "learning_rate": 2e-05,
250
+ "loss": 2.618265914916992,
251
+ "step": 310
252
+ },
253
+ {
254
+ "epoch": 0.045714285714285714,
255
+ "grad_norm": 0.044677734375,
256
+ "learning_rate": 2e-05,
257
+ "loss": 2.6028215408325197,
258
+ "step": 320
259
+ },
260
+ {
261
+ "epoch": 0.047142857142857146,
262
+ "grad_norm": 0.048828125,
263
+ "learning_rate": 2e-05,
264
+ "loss": 2.6143060684204102,
265
+ "step": 330
266
+ },
267
+ {
268
+ "epoch": 0.04857142857142857,
269
+ "grad_norm": 0.0498046875,
270
+ "learning_rate": 2e-05,
271
+ "loss": 2.602345275878906,
272
+ "step": 340
273
+ },
274
+ {
275
+ "epoch": 0.05,
276
+ "grad_norm": 0.06494140625,
277
+ "learning_rate": 2e-05,
278
+ "loss": 2.597218704223633,
279
+ "step": 350
280
+ },
281
+ {
282
+ "epoch": 0.05142857142857143,
283
+ "grad_norm": 0.07568359375,
284
+ "learning_rate": 2e-05,
285
+ "loss": 2.6153676986694334,
286
+ "step": 360
287
+ },
288
+ {
289
+ "epoch": 0.05285714285714286,
290
+ "grad_norm": 0.06787109375,
291
+ "learning_rate": 2e-05,
292
+ "loss": 2.610495948791504,
293
+ "step": 370
294
+ },
295
+ {
296
+ "epoch": 0.054285714285714284,
297
+ "grad_norm": 0.05224609375,
298
+ "learning_rate": 2e-05,
299
+ "loss": 2.6116031646728515,
300
+ "step": 380
301
+ },
302
+ {
303
+ "epoch": 0.055714285714285716,
304
+ "grad_norm": 0.08837890625,
305
+ "learning_rate": 2e-05,
306
+ "loss": 2.6053627014160154,
307
+ "step": 390
308
+ },
309
+ {
310
+ "epoch": 0.05714285714285714,
311
+ "grad_norm": 0.08447265625,
312
+ "learning_rate": 2e-05,
313
+ "loss": 2.599230194091797,
314
+ "step": 400
315
+ },
316
+ {
317
+ "epoch": 0.05714285714285714,
318
+ "eval_loss": 3.0280556678771973,
319
+ "eval_runtime": 4.3932,
320
+ "eval_samples_per_second": 67.831,
321
+ "eval_steps_per_second": 3.87,
322
+ "step": 400
323
+ },
324
+ {
325
+ "epoch": 0.05857142857142857,
326
+ "grad_norm": 0.0712890625,
327
+ "learning_rate": 2e-05,
328
+ "loss": 2.6664922714233397,
329
+ "step": 410
330
+ },
331
+ {
332
+ "epoch": 0.06,
333
+ "grad_norm": 0.07373046875,
334
+ "learning_rate": 2e-05,
335
+ "loss": 2.6177946090698243,
336
+ "step": 420
337
+ },
338
+ {
339
+ "epoch": 0.06142857142857143,
340
+ "grad_norm": 0.064453125,
341
+ "learning_rate": 2e-05,
342
+ "loss": 2.6283496856689452,
343
+ "step": 430
344
+ },
345
+ {
346
+ "epoch": 0.06285714285714286,
347
+ "grad_norm": 0.04736328125,
348
+ "learning_rate": 2e-05,
349
+ "loss": 2.6089170455932615,
350
+ "step": 440
351
+ },
352
+ {
353
+ "epoch": 0.06428571428571428,
354
+ "grad_norm": 0.06298828125,
355
+ "learning_rate": 2e-05,
356
+ "loss": 2.625172233581543,
357
+ "step": 450
358
+ },
359
+ {
360
+ "epoch": 0.06571428571428571,
361
+ "grad_norm": 0.109375,
362
+ "learning_rate": 2e-05,
363
+ "loss": 2.6342344284057617,
364
+ "step": 460
365
+ },
366
+ {
367
+ "epoch": 0.06714285714285714,
368
+ "grad_norm": 0.048583984375,
369
+ "learning_rate": 2e-05,
370
+ "loss": 2.6153539657592773,
371
+ "step": 470
372
+ },
373
+ {
374
+ "epoch": 0.06857142857142857,
375
+ "grad_norm": 0.087890625,
376
+ "learning_rate": 2e-05,
377
+ "loss": 2.5881879806518553,
378
+ "step": 480
379
+ },
380
+ {
381
+ "epoch": 0.07,
382
+ "grad_norm": 0.047607421875,
383
+ "learning_rate": 2e-05,
384
+ "loss": 2.6207881927490235,
385
+ "step": 490
386
+ },
387
+ {
388
+ "epoch": 0.07142857142857142,
389
+ "grad_norm": 0.054931640625,
390
+ "learning_rate": 2e-05,
391
+ "loss": 2.632369041442871,
392
+ "step": 500
393
+ },
394
+ {
395
+ "epoch": 0.07142857142857142,
396
+ "eval_loss": 3.028033494949341,
397
+ "eval_runtime": 4.349,
398
+ "eval_samples_per_second": 68.521,
399
+ "eval_steps_per_second": 3.909,
400
+ "step": 500
401
+ }
402
+ ],
403
+ "logging_steps": 10,
404
+ "max_steps": 7000,
405
+ "num_input_tokens_seen": 0,
406
+ "num_train_epochs": 9223372036854775807,
407
+ "save_steps": 500,
408
+ "stateful_callbacks": {
409
+ "TrainerControl": {
410
+ "args": {
411
+ "should_epoch_stop": false,
412
+ "should_evaluate": false,
413
+ "should_log": false,
414
+ "should_save": true,
415
+ "should_training_stop": false
416
+ },
417
+ "attributes": {}
418
+ }
419
+ },
420
+ "total_flos": 1.63077257428992e+17,
421
+ "train_batch_size": 18,
422
+ "trial_name": null,
423
+ "trial_params": null
424
+ }
last-checkpoint/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1ff39bd552f28407a4692ecdd5eb017e018aa993cdeeb1b9fe03f5f01952f94
3
+ size 5265